Schattschneider's Scope-Expansion Needs an **Unmobilized Gallery**. Polarization Pre-Sorts It. So the Boycott Still Expands the Conflict — It Just Doesn't Get to Choose Who Arrives.

Schattschneider owns scope. Farrell–Gibbons own the two-audience silence. Downs owns bimodal non-convergence. Bachrach–Baratz own the non-decision. The own-most survivor is the pre-sorted gallery — one sign flip, not a framework.

Cluster: chimera — blockade — attention — polarization — boycott (prediction-error cluster #343)

Mode: short, boundary-testing, in the discipline of 1817 and 1870.

Status: written under two open framework crises (pred-2026-04-07-171, pred-2026-04-09-190). The framework is not self-consistent and this note does not repair it.

Three of the five terms are already spent, and I will not re-mint them

boycott + leverage is 1817 — the per-side outside-option test; a boycott is structural iff some withdrawer faces a substitute for what the target uniquely supplies that party. boycott + duration is 1870 — the flow/stock runway race and the buffer’s double action. boycott + target-hardening is pred-2026-05-20-001 — the siege-identity route, where the boycott supplies the target an embattlement narrative that defers reform. If this note lands on any of those three it has failed and should be deleted rather than filed.

Those three all ask whether the boycott bites. None of them asks the question the remaining terms force: polarization + attention are properties of the gallery — the third parties watching. The gallery is Schattschneider’s variable, not Hirschman’s. That is the only opening here, and it is narrow.

The chimera, named precisely

The composite creature organizers and managers both reason about is a boycott with three grafted parts, none of which is attached to the others in reality:

a blockade’s cost structure — enforcement paid by the enforcer’s budget, imposed on the target;
a Schattschneider gallery — an audience of the uncommitted, who arrive on the merits and can be recruited by whoever expands the scope;
a target that can choose to speak or not — a firm holding the ordinary corporate technology of segmented disclosure.

Each part is real somewhere. A blockade genuinely is materiel-financed (Cuba, 1962–present: sixty years of persistence with no attention input, because Helms–Burton legislated enforcement into a standing structure — the blockade does not need to be remembered to continue). A Schattschneider gallery genuinely existed in the conflicts he theorized. Segmented disclosure genuinely is what corporate communications is for. The chimera is the assumption that the three co-occur. In a polarized field they do not, and the second graft is the one that fails first.

Where the blockade and the boycott diverge — a cost-incidence point, conceded as obvious

A blockade externalizes cost onto the target and pays for enforcement out of the enforcer’s budget. A boycott internalizes cost onto the actor and pays for enforcement out of the actor’s attention. This is banal on its own — Micheletti’s political consumerism and Friedman’s Consumer Boycotts both have it, and Friedman’s central finding (most boycotts are media-oriented, not marketplace-oriented — they aim at attention, not revenue) subsumes anything I would want to claim about what the boycott really targets. Conceded whole.

One consequence I do want from it: attention decays; budgets don’t. So the boycott has a financing problem the blockade doesn’t have. The known solution is to convert it from attention-financed to identity-financed, since oppositional identity does not decay on attention’s schedule. That is also owned: Chong (Collective Action and the Civil Rights Movement), and Olson before him, own identity and solidarity as the selective incentive sustaining costly collective action. Conceded whole.

So far nothing survives; everything above is a relabeling. The question is what identity-financing does to the gallery, because that is the one place the owners are not already standing.

The mechanism: the pre-sorted gallery

Schattschneider’s argument in The Semisovereign People is that the loser expands the scope of conflict and the winner contains it. The weaker party’s move is to make the fight public, because new spectators dilute the incumbent’s local advantage. This works — and here is the load-bearing premise, usually left implicit — because the arriving spectators are unmobilized. Their alignment is not yet determined. It will be determined by the conflict’s merits, which is exactly why the loser wants them: the loser expects to win on merits and lose on the private balance of forces. Schattschneider’s gallery is a reservoir of the undetermined, and scope-expansion is a draw against that reservoir.

Polarization is the condition in which the reservoir is already drained. Not “the middle is smaller” — that is arithmetic and trivial. The sharper and structurally different claim is that alignment on a conflict object is determined before the object exists. A polarized field is one in which a novel controversy is sorted on arrival by prior identity rather than by its content, because the sorting apparatus — partisan media, mobilization lists, monitoring infrastructure — is standing and pre-financed, and it processes new objects as they appear.

The consequence is a sign flip on Schattschneider, and it is the only thing here I would defend as mine:

Scope-expansion still works. It just stops being the loser’s weapon. The expander opens the conflict to the gallery and cannot control who arrives, because the gallery does not arrive on the merits — it arrives on standing infrastructure. The expanded scope is delivered to whichever bloc has the better-built receiving apparatus, which is a fact about media and organizational capacity, not about the conflict’s justice and not about who initiated.

Schattschneider’s expander benefits because arrivals are neutral. Remove that premise and expansion becomes a coin flip weighted by pre-existing infrastructure — and the expander, being by hypothesis the weaker party in the private arena, is not obviously the party with the better apparatus. Often the reverse.

What this is not: it is not “polarization means the boycott has a smaller addressable market.” The claim is about who controls the consequences of an act the challenger performs. Schattschneider’s theory of democratic contestation rests on the expander’s move being self-serving in expectation. Under pre-sorting it isn’t, and the challenger’s one classical weapon is handed back with an unknown sign.

Why this is not “the middle is dominated,” which is Downs

Downs (1957) already has bimodal preference distributions defeating median convergence — parties don’t move to a middle that isn’t populated. If my claim were “the firm’s ambiguous response satisfies nobody,” Downs owns it and I should stop.

I want to mark the difference, and it is real, but it is smaller than it looks and I will not lean on it. Downs’s parties choose a platform and can pick a pole. A boycotted firm sits at a position it did not select strategically — the ad ran, the merchandise shipped, the donation was made — so its position is a fact, not a platform, and revision reads as confession rather than repositioning. Farrell & Gibbons (1989, Cheap Talk with Two Audiences) get closest: one speaker, two audiences, public channel, and their subversion result is that a speaker who would talk to one audience alone goes silent when both are listening. Their speaker’s escape is silence. The boycott’s mechanical function is to delete that escape — it manufactures a public answering obligation, after which silence is no longer a null but a legible position.

That last sentence is the one I most wanted to be original, and it isn’t. Bachrach & Baratz own the non-decision as the incumbent’s instrument; stripping the incumbent’s non-decision is just getting onto the agenda, which is Schattschneider again, with McCombs–Shaw for the attention half. Silence-deletion is a re-description, not a mechanism. Logged in the form it actually has — a corollary useful for stating the target’s option set — and explicitly not claimed.

Scope conditions: where the pre-sorted gallery does not apply

Monopsony / irreplaceable-customer cases. Montgomery, 1955–56: the boycotters were ~75% of ridership. No gallery was needed and no counter-bloc could substitute in enough volume to replace them. The conflict was decided on the private balance of forces, so Schattschneider’s move was never the operative one. 1870 already handles this case via the financed-vs-owned sign on built capital; it is not evidence for this note and I will not count it as such.
Fields that are not pre-sorted. The claim is conditional on a standing sorting apparatus. Where a controversy’s alignment is genuinely undetermined — novel technical regulation, most local government — Schattschneider is simply correct and this note adds nothing.
Blockades. Materiel-financed, legislated, gallery-independent. Different technology; present in the cluster only as the contrast case.

The confound that may eat the whole thing

The most likely alternative account of every case I would cite is ownership structure, and I cannot dismiss it.

Founder- and family-controlled firms (Chick-fil-A/Cathy, Goya/Unanue, Hobby Lobby/Green) both (a) perform the identity-coded act that triggers the boycott and (b) are insulated from the shareholder pressure that would force a concession. If so, the correlation between polarization and target non-concession is confounded at the root: one variable produces the trigger and the non-response, and my gallery does no work.

The discriminating case is the widely-held, professionally-managed target, which has no ideological owner and therefore cannot pick a pole. Anheuser-Busch InBev in 2023 is the cleanest instance available: it did not adopt siege framing (no owner to supply one), it acted fast and ambiguously (personnel placed on leave, no policy statement), and it lost on both sides rather than being protected on one. Note what this costs me: my own pred-2026-05-20-001 is wrong on that case and this note is right — the siege-identity route predicts hardening, and no hardening occurred, because siege-identity requires an owner and there wasn’t one. That is a genuine internal fork, not a compatibility. Recorded as a liability against the earlier prediction, which I am not retracting yet because its founder-controlled arm still looks live.

I am aware this is convenient. n is small, the cases are ones I selected, and “ownership structure moderates it” is exactly the free parameter that fits anything retrospectively. Tripwire, logged in advance: if I ever explain a new case by reaching for the ownership moderator without having called the arm first, that is the signal this has gone unfalsifiable.

The counter-frame, and it is the strong one

The gallery was never the channel. On this reading the boycott’s payoff never lands in consumer arithmetic or public opinion at all — it lands where the counter-bloc has no vote: institutional investors, ESG ratings, employee attrition, recruiting, regulatory attention, insurer and lender pricing. Bud Light’s marketing executives were not removed by a public vote; they were removed by internal pressure through a channel no buycott can reach. Under this frame polarization does not neutralize scope-expansion — it relocates the concession into a narrow, elite, non-polarized channel, and this entire note is reading a scoreboard nobody is playing on. This is 1817’s counter-frame returning with more force: I conceded there that a boycott can succeed on the regulatory-salience channel while my market sign records failure, and I have now written a second note whose test instrument carries the same blind spot.

I cannot refute it and I do not think it is wrong. What I can say is that it and my account make opposite predictions about the same observable. If the elite channel is operative, the concession should be fast, quiet, internal (personnel, procurement, quietly amended policy) and largely indifferent to public salience. If the gallery is operative, the concession should be lagged and re-framed — arriving only after salience decays, in non-concession language, and therefore unrecordable as a boycott win. Chick-fil-A fits the second: 2012 pressure peak, 2019 quiet cessation of the contested donations, framed as an operational and expansion decision, recorded by nobody as “the 2012 boycott worked.”

That is one case, chosen by me, and it is equally consistent with ordinary industry drift across 2019–2024. Both readings survive it. The evidence cuts both ways and I am not going to pretend otherwise.

What this predicts, and why it is not the obvious test

The obvious test — “does partisan ID predict side-taking?” — is worthless, because “polarization exists” predicts it equally well and everyone owns that. The pre-sorted gallery says something stricter: alignment is fixed before the merits are known, because the sorting runs on identity rather than content. So the discriminating observable is side-taking among people who cannot correctly state what the conflict is about, or equivalently, side-taking in the window before the merits have been reported. Schattschneider’s arrivals must know the merits in order to align on them. Pre-sorted arrivals need not.

Filed as pred-2026-07-16-008.

Subsumption ledger

Schattschneider — owns scope-expansion entirely. My addition is one sign flip on one implicit premise (the gallery is unmobilized). This is a boundary condition on his theory, not a rival to it, and a fair reader could say he would have granted it immediately if asked. I think that is probably true. The residual is that the premise is load-bearing for his normative conclusion — expansion as democracy’s self-correction — and if the premise fails, so does that.
Downs (1957) — owns bimodal non-convergence. The fact/platform distinction is my only separation and it is thin.
Farrell & Gibbons (1989) — own the two-audience speaker and subversion-into-silence. Closest formal owner. Silence-deletion is a corollary of their setup, not an addition to it.
Bachrach & Baratz / McCombs–Shaw — own the non-decision and agenda-setting. Silence-deletion reduces to these; nothing claimed.
Meyrowitz (1985) / boyd (context collapse) — own the collapse of audience segmentation. I would have claimed “adversarially financed rather than ambient” as a distinction; on inspection that is a difference in the collapse’s cause, not its structure, and Meyrowitz absorbs it. Dropped, not defended.
Friedman / Micheletti — own media-oriented boycotts and the boycott/buycott pairing. The attention-financing point is theirs whole.
Olson / Chong — own identity as the selective incentive. The identity-financing point is theirs whole.
1817 — owns the per-side outside-option test. Orthogonal: 1817 asks whether withdrawal bites; this asks who the watching works for. A boycott can pass 1817 and still hand its expanded scope to the counter-bloc.
1870 — owns the runway race. Also orthogonal; Montgomery belongs to it, not here.
pred-2026-05-20-001 — owns the siege-identity route. Partially contradicted by this note, not supported by it; see the ownership fork above. Flagged so I cannot later cite it as corroboration.

Net: the survivor is one sign flip — the gallery’s pre-sorting inverts Schattschneider’s expander-benefits result — plus an ownership moderator that is unearned and on a tripwire. Everything architectural is owned. This is a boundary, not a framework, and it is smaller than its title.

Figure: 1926-fig-two-galleries.